This thesis presents a comprehensive approach to enhancing camera-based optical wireless communication by addressing the inherent limitations of conventional frame-based visible light communication (VLC) systems, such as low frame rates, motion blur, and high latency. The research is structured around two primary dimensions—spatial and temporal diversity—to significantly improve data rate, reliability, and user experience.

In the first part, we introduce Revelio, a novel screen-camera communication system that transforms every pixel of a display into an individual LED. By employing spatially-adaptive flicker fusion techniques and encoding data in the shapes of pixel regions rather than traditional Manchester-style methods, Revelio achieves imperceptible data embedding with efficient symbol packing and robust decoding. The system leverages neural networks for both frame identification and symbol detection, and its performance is validated using industry-standard video quality metrics under various operational conditions.

The second part enhances the imperceptibility of embedded codes through advanced optimization strategies. A Differential Evolution optimizer is employed to address the constraints of the encoding process, complemented by a novel Generative Adversarial Optimizer that uses a RandomForest model to select optimal modulation parameters. This dual-optimizer framework maintains high decodability while ensuring minimal visual artifacts.

In the third part, the integration of neuromorphic cameras—capable of sampling at up to 20 kHz with high spatial resolution—pushes the temporal boundaries of data capture. These cameras not only support rapid data transmission through an adaptive n-pulse encoding technique (achieving error rates in the order of $10^{-3}$) but also enable joint sensing and communication, demonstrated via an event segmentation self-supervised model for object detection and obstacle avoidance.

Collectively, this thesis advances the field of optical wireless communication by seamlessly integrating spatial and temporal innovations, thereby opening new avenues for high-speed, reliable, and energy-efficient wireless systems with integrated sensing capabilities.